In this investigation, the effect of various load histories on the fatigue behaviour of plain concrete was studied. (1). A total of 462 cylindrical specimens (100x250 mm) were tested statically and dynamically in compression. In addition, a total of 18 cylinders (150x300 mm) plus 132 cubes (100 mm) were tested statically. In addition to fatigue strength, special interest has been paid to the deformation characteristics of the concrete. A total of 140 specimens were tested under constant amplitude loading. Based on the fatigue strength results, empirical ex-pressions between the stress level (S), number of cycles (N) and probability of failure (P), S-N-P relationships, were derived. Using deformation characteristic results, an empirical expression for the total longitudinal strain (&maX) as a function of the cycle ratio (N/NF) and the loading time (t) was derived. This expression can be used to predict the fatigue life from deforma-tions early in the life. A total of 180 specimens were tested under various variable load histories of a given statistical distribution. The effect of both small and large amplitudes on fatigue strength and deforma-tion characteristics was examined. Based on fatigue strength results, the validity and limitations of the Palmgren-Miner hypothesis were examined. The PM hypothe-sis was found to give more or less unsafe predictions of the fatigue life depending on the load histories. Features of the loading histories affecting the accuracy of the hypothesis are discussed and an empirical relationship between loading histogram parameters and the Miner-sum at failure is presented. The empirical expression for the total longitudinal strain de-rived for constant amplitude tests is modified for tests under variable load histories.

SP-75 The eighteen papers included in this volume report on international research into the fatigue of concrete structures. Among the topics are fatigue in structures subject to cyclic loading in offshore and Arctic environments; hydraulic fracturing effects of water; marine corrosion and fatigue strength; the validity of Miner's hypothesis; and methods of predicting crack widths and fatigue loading.

The results show that (a) the flexural fatigue life of normal concrete beams is higher than the fatigue life of layered beams utilizing dense concrete or latex modified concrete at lower stress levels but is lower than that of hybrid layered beams at higher stress levels and (b) essentially the same fatigue strength was obtained for beams containing only a 2-inch layer of either dense concrete or latex modified concrete at the surface as for beams containing these concretes full depth.

As part of an ongoing experimental/analytical research effort to evaluate the feasibility of a test method for fracture toughness of concrete, forty-eight plain concrete beams have been tested in bending to failure. All beams were notched and then precracked to different crack length/depth ratios-prior to load-ing to failure. The precracking was done using an electro-hydrodynamic materials testing system and displacement control. The beams which were cracked in fatigue were subjected to one million cycles of sinusoidal loading at 4 Hz. After the cycling was complete on a beam, the crack depth was determined using a compliance calibration technique following which the beam was loaded to failure. A load versus crack-mouth-opening-displacement trace was plotted during this final load run. For each beam tested in fatigue, a companion beam was pre-cracked "statically" by loading in repeated cycles until the crack depth, as measured by compliance calibration, matched that of the fatigued specimen. The studies were made on two different beam sizes in three and four-point bending with two different mix designs. Test results indicate the failure strength and associated maximum stress-intensity of the statically precracked beams to be slightly higher than those precracked in fatigue.

The objective of this investigation was to study ex-perimentally the behavior of plain concrete subjected to slow cyclical loading in compressive uniaxial and biaxial states of stress. A triaxial testing machine which had been successfully used to study the response of plain concrete cubicai specimens subjected to static multiaxial stress states was used in this study. The test specimens were subjected to prescribed stress histories which included 15 maximum stress levels. In addition to the uniaxial state of stress, biaxial stress states of two types were studied. The first was a proportional loading type in which two loading paths were used, namely 02/q = I .O and 02/01 = 0.5. The second loading type consisted of a constant stressing the direction with a cyclical loading in the 01 direction. The cyclical loading ranged from zero to various percentages of the unconfined static strength of the cube. Cycle rates used were in the range of one- cycle per minute. Strain measurements in all three principal directions were recorded for each test. This data was used to show the effects of number of cycles, load paths, and state of stress on the dilatational, as well as stress-strain response for plain concrete. The data waslso used to show the effect of state of stress and load path on the fatigue strength of concrete.